Aug 30, 2025 · Aug 30, 2025 · Aug 30, 2025 · Aug 30, 2025 · Aug 31, 2025
diff --git a/coderd/prometheusmetrics/aggregator_test.go b/coderd/prometheusmetrics/aggregator_test.go
 package prometheusmetrics_test

 import (
 "bytes"
 "context"
 "fmt"
 "sort"
 "github.com/stretchr/testify/assert"
 "github.com/stretchr/testify/require"

 "cdr.dev/slog"
 "cdr.dev/slog/sloggers/sloghuman"
 "cdr.dev/slog/sloggers/slogtest"
 "github.com/coder/coder/v2/coderd/agentmetrics"

 agentproto "github.com/coder/coder/v2/agent/proto"
 "github.com/coder/coder/v2/coderd/agentmetrics"
 "github.com/coder/coder/v2/coderd/prometheusmetrics"
 "github.com/coder/coder/v2/cryptorand"
 "github.com/coder/coder/v2/testutil"

 // given
 registry := prometheus.NewRegistry()
 metricsAggregator, err := prometheusmetrics.NewMetricsAggregator(slogtest.Make(t, &slogtest.Options{IgnoreErrors: true}), registry, time.Hour, tc.aggregateOn) // time.Hour, so metrics won't expire

 logger := slogtest.Make(t, &slogtest.Options{IgnoreErrors: true})
 logger = logger.Leveled(slog.LevelDebug)

 // updateWatch is how we know the update was processed before we collect
 var updateWatch bytes.Buffer
 logger = logger.AppendSinks(sloghuman.Sink(&updateWatch))

 metricsAggregator, err := prometheusmetrics.NewMetricsAggregator(logger, registry, time.Hour, tc.aggregateOn) // time.Hour, so metrics won't expire
 require.NoError(t, err)

 ctx, cancelFunc := context.WithCancel(context.Background())
 metricsAggregator.Update(ctx, sc.labels, sc.metrics)
 }

 // This code makes sure all updates are processed before we start a collection.
 // We do this by grepping the logs for "update metrics" messages, which are logged
 // every time an update is processed.
 //
 // This is not ideal as the code relies on implementation details (the log message),
 require.Eventually(t, func() bool {
 return strings.Count(updateWatch.String(), "update metrics") == len(tc.given)
 }, testutil.WaitMedium, testutil.IntervalFast)

 // then
 require.Eventually(t, func() bool {
 var actual []prometheus.Metric
Original file line number	Diff line number	Diff line change
		@@ -1,6 +1,7 @@
		package prometheusmetrics_test

		import (
		"bytes"
		"context"
		"fmt"
		"sort"
Expand All		@@ -14,10 +15,11 @@ import (
		"github.com/stretchr/testify/assert"
		"github.com/stretchr/testify/require"

		"cdr.dev/slog"
		"cdr.dev/slog/sloggers/sloghuman"
		"cdr.dev/slog/sloggers/slogtest"
		"github.com/coder/coder/v2/coderd/agentmetrics"

		agentproto "github.com/coder/coder/v2/agent/proto"
		"github.com/coder/coder/v2/coderd/agentmetrics"
		"github.com/coder/coder/v2/coderd/prometheusmetrics"
		"github.com/coder/coder/v2/cryptorand"
		"github.com/coder/coder/v2/testutil"
Expand DownExpand Up		@@ -592,7 +594,15 @@ func TestLabelsAggregation(t *testing.T) {

		// given
		registry := prometheus.NewRegistry()
		metricsAggregator, err := prometheusmetrics.NewMetricsAggregator(slogtest.Make(t, &slogtest.Options{IgnoreErrors: true}), registry, time.Hour, tc.aggregateOn) // time.Hour, so metrics won't expire

		logger := slogtest.Make(t, &slogtest.Options{IgnoreErrors: true})
		logger = logger.Leveled(slog.LevelDebug)

		// updateWatch is how we know the update was processed before we collect
		var updateWatch bytes.Buffer
		logger = logger.AppendSinks(sloghuman.Sink(&updateWatch))

		metricsAggregator, err := prometheusmetrics.NewMetricsAggregator(logger, registry, time.Hour, tc.aggregateOn) // time.Hour, so metrics won't expire
		require.NoError(t, err)

		ctx, cancelFunc := context.WithCancel(context.Background())
Expand All		@@ -606,6 +616,15 @@ func TestLabelsAggregation(t *testing.T) {
		metricsAggregator.Update(ctx, sc.labels, sc.metrics)
		}

		// This code makes sure all updates are processed before we start a collection.
		// We do this by grepping the logs for "update metrics" messages, which are logged
		// every time an update is processed.
		//
		// This is not ideal as the code relies on implementation details (the log message),
		require.Eventually(t, func() bool {
		return strings.Count(updateWatch.String(), "update metrics") == len(tc.given)
		}, testutil.WaitMedium, testutil.IntervalFast)
Comment on lines +619 to +626 Copy link MemberAuthor EmyrkAug 30, 2025• edited Loading Choose a reason for hiding this comment The reason will be displayed to describe this comment to others.Learn more. This is a really unfortunate fix. There is no way to hook into the timing. Maybe with quartz? The issue is`time.Now()` is called before the channel enqueue. So to do this with quartz, we would need to add some other`time.Now()` function in the update select case. Copy link Member johnstcnAug 31, 2025 Choose a reason for hiding this comment The reason will be displayed to describe this comment to others.Learn more. This may need to be some kind of buffer protected by a mutex, or perhaps a buffered channel instead?

		// then
		require.Eventually(t, func() bool {
		var actual []prometheus.Metric
Expand Down